Missing Parallels

There are many parallels among the Pāli that aren’t marked as such on Sutta Central. Bhante @sujato - Do you want us to report these? If so: where? how?

Indeed, we haven’t updated our parallels in a long time, and I would love to improve them. Problem is, there is no person doing this work, and no regular way of making updates, except by editing the JSON files by hand.

So … nice of you to volunteer?

1 Like

Happy to take it. I’m familiar with the “parallels methodology” already.

  1. Is there any other documentation or guidance on what to count as which kind of parallel?
  2. Which JSON is the right one to edit?
  3. Is there an existing list of parallels desired to be added? Or just open Pull Requests as I find them?



Awesome, I was only half joking!

Okay, so this task has lain fallow for some time, so there’s no real plan or system ATM.

Not really, the general idea is that it should be considered a parallel by relevant scholars. Since the idea is, “if you’re studying this, that might worth taking a look at too”, it’s better to err on the side of false positives rather than being too critical.

This work is, in my view, orthoganal to something like Buddhanexus, which relies on ML to find parallels. We should aim to represent the findings of scholarship.

The parallels are all in one big JSON file.

The file structure should be clear enough, tho I’m not sure this is the best way to do it. There’s also some logic involved, for example, if a sutta is a full parallel, the relation works both ways. So you probably want to check a few cases and see how they present on SC.

  • basic refs use sutta IDs.
  • hash refers to internal links
  • ~ prefix is a “resembling parallel”
  • mentions are when one sutta “mentions” another by name.

One thing to bear in mind, this system pre-dates Bilara. So it uses hash whereas Bilara uses colon and converts to hask for the URL.

More importantly, the indicated section, especially for Pali texts, is not necessarily (or perhaps ever) the actual Bilara segment ID. I’ve wanted to sort this out for ages! However, the Bilara segment IDs are based on the pre-existing paragraph IDs used here, so in most cases it should not be difficult to correct, but I haven’t really investigated the scope of the problem.

This is briefly noted in this issue:

And I just did a bit of an investigation.

There is a list of outstanding issues.

The two things I’ve always wanted most would be:

  • add parallels for snp, these are in BB’s translation.
  • review mn parallels with Analayo’s Comparative Study.


"mentions": ["an4.34","~ne37#511","~mil5.5.3#8"]

I assume the semantics of this is “an4.34” is mentioned by ne37 and mil5.5.3 at spots 511 and 8 and the ~ means “not by name, but basically it’s clear which sutta they’re referring to”?

1 Like

Where can I learn more about this?

Great. I noticed the issue when I was going through Hecker’s Similes so I could add that one to the list.